Researchers automate brain MRI image labelling, more than 100,000 exams labelled in under 30 minutes

Researchers automate brain MRI image labelling, more than 100,000 exams labelled in under 30 minutes

Researchers from the School of Biomedical Engineering & Imaging Sciences at King’s College London have automated brain MRI image labeling, needed to teach machine learning image recognition models, by deriving important labels from radiology reports and accurately assigning them to the corresponding MRI examinations. They can now label more than 100,00 MRI examinations in less than half an hour.

The further challenges for them will be to perform the deep learning image recognition tasks to ensure the developed models can still perform accurately across different hospitals using different scanners.

Read the full article: https://www.eurekalert.org/pub_releases/2021-07/kcl-rab072221.php

A noninvasive test to detect cancer cells and pinpoint their location

Most of the tests that doctors use to diagnose cancer — such as mammography, colonoscopy, and CT scans — are based on imaging. More recently, researchers have also developed molecular diagnostics that can detect specific cancer-associated molecules that circulate in bodily fluids like blood or urine.

MIT engineers have now created a new diagnostic nanoparticle that combines both of these features: It can reveal the presence of cancerous proteins through a urine test, and it functions as an imaging agent, pinpointing the tumor location.  The diagnostic could be used to monitor the progression of colon cancer, including the spread of metastatic tumors to the lung and the liver. Eventually it could be developed into a routine cancer test that could be performed annually.

Read the full article : https://www.nanowerk.com/nanotechnology-news2/newsid=58460.php

A new model for radiology virtual consultations

A group of U.S. researchers found that virtual consultations can improve radiology’s value by enabling face-to-face visits among radiologists, primary care physicians (PCPs), and their patients.

In the study participating radiologists could accept consultations through a video- and microphone-enabled PACS station in the hospital. Moreover, the visits lasted just eight minutes each on average. The radiologist discussed the imaging findings and answered any questions about the findings or follow-up recommendations with the patient and the PCP.

Their study demonstrates the feasibility of point-of-care virtual radiology consultations in primary care settings and provides preliminary data on the benefit of such a practice mode.

Read the full article : https://www.auntminnie.com/index.aspx?sec=ser&sub=def&pag=dis&ItemID=133006

Cutting edge of medical imaging in India

There have been rapid advancements in imaging techniques being used in India. Two of these are cryogenic electron microscopy (cryo-EM) and photoacoustic imaging.

Indian government’s science research funding agency announced that India would get four new cryo-EMs. Cryo-EM is a protein imaging technique. It will help medical professionals give quicker detection and response to any new disease.

Meanwhile Photoacoustic systems are built on the conventional ultrasound imaging platform; the only additional cost is of the light source. As such, the cost to a patient is comparable to an ultrasound scan. This artificial intelligence-augmented photoacoustic imaging technique is an advancement in cancer diagnostics.

Read the full article: https://www.thehindubusinessline.com/business-tech/cutting-edge-of-medical-imaging/article34866758.ece

Deep-learning model improves radiologist interpretation of X-rays

Evidence suggests that deep-learning systems show great potential for the detection of lesions and pattern classification on chest radiographs. Scientists did comparative research for both the radiologists and a deep-learning system. 

In their discussion, the authors suggested that the study showed the potential for deep-learning to improve chest X-ray interpretation across a wide range of clinical findings and concluded that further work is underway to confirm the applicability of the model as an efficient diagnostic tool in clinical practice as opposed to a substitute for radiologists.

Read the full article:

https://hospitalhealthcare.com/clinical/radiology-and-imaging/deep-learning-model-improves-radiologist-interpretation-of-x-rays/

Thermal imaging predicts venous leg ulcer healing

A team of researchers from Australia determined that thermal imaging can tell a provider within two weeks of an initial scan if a leg ulcer will heal within 12 weeks. This technique could replace the current digital strategies or invasive planimetric tracing. A non-contact method like thermal imaging would be ideal to use when managing wounds in the home setting to minimize physical contact and, therefore, reduce infection risk.

This technique accurately measures changes in wound size and other physiological parameters over time. It is also a more time-efficient and cost-effective method.

Read the full article:

https://www.diagnosticimaging.com/view/thermal-imaging-predicts-venous-leg-ulcer-healing

FDA approves first PSMA PET imaging agent for prostate cancer

US FDA approves the Piflufolastat F-18 injection for prostate cancer detection It is the first fluorinated prostate-specific membrane antigen (PSMA) agent approved by the FDA commercially available PSMA PET imaging agent. It allows doctors to see suspected prostate cancer at an early stage.

This is a highly effective imaging approach to detect the spread of cancer to other parts of the body.

Read the full article: 

https://www.itnonline.com/content/fda-approves-first-commercially-available-psma-pet-imaging-agent-prostate-cancer

Korea makes medical imaging smarter with big data & deep learning

A research team from the Korean Ministry of Science has developed a technology that improves both the speed and accuracy of disease diagnosis. They applied big data deep learning technology that has been used to diagnose the reliability of mechanical parts and equipment to ultrasound imaging equipment to develop diagnostic imaging assistance technology using machine learning.

The research team plans to modify the deep learning model to improve the accuracy of aortic plaque analysis. This technology achieves a diagnosis with an accuracy of 80%.

Read the full article :

https://www.biospectrumasia.com/news/51/18342/korea-makes-medical-imaging-smarter-with-big-data-deep-learning.html